Another major factor is that we didn’t just get the engine, we got the entire code base, which saved a massive amount of time / money because we didn’t have to recreate an entire ARPG gameplay, feature set and tools from scratch.
Even if you license a commercial engine, you still have a ton of work to do before you have any semblance of a game - unless what you’re making is very similar to what that engine was created to do. If you’re using Unreal to make a Gears of War clone, then you’d already have native support for a lot of the gameplay you wanted to create. If you’re making an ARPG, you’re not starting with much besides and engine and then you have to sift through or even try to rework existing code that is irrelevant to your game or not suited for it. Even the way levels are made in Unreal is not great for the way we wanted to build GD, so we’d either have to figure out how to work with it or create a new level editor and terrain system.
We’d still have to do a bunch of work before we could even click on the ground and make a character run around in an isometric view. Then you need to add loot mechanics, player inventory, skills, a skill system, combat mechanics and formulas… basically all the things that make an ARPG.
Starting with the TQ engine, we have everything, in terms of programming, we need to make a game with all the gameplay, features and UI of TQ. Then it’s just a question of what we want to improve, add or remove. We’ve reworked huge portions of both the engine and game. On the engine side, we recently rewrote the renderer from scratch and previously replaced the physics engine, pathing engine, shader model, sound engine, added post effects, etc. The only thing really limiting us from replacing all of the engine components over time is that it has to support the art we’ve already created and we’re sort of obligated to support older versions of directX because we supported them with the initial release and people running older systems wouldn’t be able to play if we didn’t.
Another issue is that a lot of what people interpret as being the look of the engine is actually the look of the art itself and how it was made. We started working on GD 8 years ago now and expectations for the size of texture resolutions and triangles on models was lower than they are today. A lot of what has improved with the look of the game over the years is just us updating art / redoing textures.
The thing that might have the biggest visual impact now would be to redo lighting and the way materials are rendered but that isn’t really feasible unless we move on to GD2 and leave all this current art behind. If we go on to GD2 and no longer have to support older DX version or existing art, then we’d be free to make bigger updates to the engine.
Like Kamil said, the date an engine was originally created is sort of irrelevant since you can continually update it and eventually replace every part of it.